Microsoft Reduces Reliance on OpenAI, Anthropic AI Models for Productivity Apps

Microsoft is reported to have begun reducing its dependency on artificial intelligence (AI) models owned by OpenAI and Anthropic by increasing the use of its own proprietary AI models for several of its productivity applications.

According to a TechCrunch report on Tuesday (7/7) local time, Microsoft is now utilising its own AI model, known as MAI, to handle certain tasks within the Word and Excel applications. Previously, Microsoft openly stated that various AI features in Office 3-65 services were powered by models developed by OpenAI and Anthropic.

While still utilising AI technology from other tech firms, Microsoft is said to be becoming increasingly active in developing its own AI ecosystem. At the annual Build conference held last month, Microsoft introduced seven new MAI models, including models designed to assist with AI-based programming and generating images from text prompts.

Microsoft’s effort to reduce the use of third-party AI models is described as part of a broader trend within the technology industry. Following a period of aggressive AI development expansion earlier this year, several companies are now becoming more cautious in managing expenditures related to the operation of such technology.

Several major corporations, including Amazon, Uber, Meta, and Accenture, have also reportedly been limiting spending on AI technology usage. The high cost of developing and operating AI is cited as an issue gaining increasing attention within the tech industry.

Developing AI models requires massive computational resources, which directly impacts the rising operational costs of companies. Furthermore, some Silicon Valley firms are reportedly considering the use of Chinese AI models as a cheaper alternative for AI agent needs, although such a move also raises concerns regarding security and data protection.